The aim of the dissertation is to shed light on the question of what linguistic knowledge available to children language learners is innate as opposed to being gleaned post-natally from the environment, and to do it by answering a more basic question: what, if it were innate, would assist language learning, and what would not? The method involves implementing a formal model for learning a basic but quite versatile kind of grammar for natural language syntax using methods borrowed from machine learning and statistics, along with methods from theories of syntax and semantics in linguistics. Experiments comparing results from the methods to be implemented with those of other, existing unsupervised learning systems for grammatical inference as benchmarks will be carried out, computing measures of performance commonly accepted by computational linguists to assess the accuracy of competing grammars. Initially the testing will be done on small sets of data, but with a view to eventual scaling up so that the system can be trained on large sets of data characterizing actual natural language use in conversational contexts. [unreadable] [unreadable]